Boiling river in Peruvian Amazon reveals geothermal mystery without volcanic source

In the Peruvian Amazon, a tributary runs so hot that a frog falling into it cooks to death in seconds. The water reaches 86 degrees Celsius along a six-kilometre stretch, with peak temperatures exceeding 90 degrees—hotter than a commercial dishwasher's rinse cycle, as hot as the water that leaves a kettle. The locals call it Shanay-Timpishka, which translates roughly to boiled with the heat of the sun. The nearest active volcano sits more than 700 kilometres away.

For most of the twentieth century, only the Asháninka communities and a handful of oil prospectors knew about the river. Western science ignored it until the 2010s, when geothermal scientist Andrés Ruzo—whose grandfather had told him stories about it as a child—began measuring it in earnest. What he found was not volcanic. The mechanism is geological plumbing. Rainwater falling on the Andes seeps down through cracks in rock, travels several kilometres underground, warms against hot stone at depth through the ordinary geothermal gradient that exists everywhere on Earth, and rises back up through a system of faults into a tributary that eventually joins the Amazon. No magma chamber. No volcanic vent. Just water taking the long route home.

To heat water from ambient temperature to over 90 degrees without volcanic assistance requires it to reach a depth of roughly two to three kilometres and spend long enough there to equilibrate with surrounding rock. Two kilometres is deeper than most mining operations. It is the depth at which temperature stops being a curiosity and starts being a hazard. The water emerging from the ground is old—rainwater that may have spent decades or centuries underground, fed in part by rain that fell before the invention of the transistor. The heat is not being generated in real time. It is being extracted from rock that has been quietly warm for millions of years.

In 2022 and 2023, ecologists Riley Fortier of the University of Miami and Alyssa Kullberg of EPFL in Lausanne placed 13 temperature-logging devices along a two-kilometre transect running from cooler forest into the hottest stretches. They logged a year of air temperature data. The averages ranged from 24 to 25 degrees Celsius in the cooler forest to 28 to 29 degrees in the warmest zones. The hottest air readings, taken close to the steaming water, approached 45 degrees Celsius. Fortier described the fieldwork bluntly: it was like doing fieldwork in a sauna.

What they discovered was a sharp ecological pattern. For every one degree Celsius rise in average temperature, biodiversity dropped by roughly 11 per cent. Large evergreens such as Guarea grandifolia, which can grow to 50 metres, struggled near the hottest parts. Understorey vegetation thinned out. Leaf litter turned crunchier and drier. The canopy became patchier. Not every plant suffered. The giant Ceiba tree, Ceiba lupuna, which stores water in its swollen trunk, proved unusually resilient—the botanical equivalent of a camel. Small drought-tolerant species crept in as the larger trees dropped out. The forest did not disappear. It simplified. The whole transition from ordinary Amazon forest to something scrubbier and drier happens over a stretch shorter than an easy afternoon walk.

The reason plant ecologists care about Shanay-Timpishka is that it is a natural experiment nobody could design. You cannot artificially heat a section of Amazon rainforest to see how it responds to warming. You can only find a place where nature has already done it. The direction of the effect—biodiversity thinning, canopy simplifying, drought-tolerant species advancing—matches predictions from broader tipping-point studies. Chris Boulton of the University of Exeter, a co-author on the 2023 global tipping points report, called the natural-experiment framing a clever thing to do. But the Amazon spans more than 6.7 million square kilometres across nine countries. One two-kilometre stretch cannot stand in for all of it. The steam alone makes the local microclimate unusual, and rainfall patterns across the basin will shift in ways this one tributary cannot capture.

The exact plumbing under Shanay-Timpishka is still being mapped. The fault geometry, the recharge zones, the residence time of the water underground—these are being estimated rather than measured directly. The ecology paper is more limited still. It measures air temperature and plant composition. It does not measure insect populations, though Fortier suspects the steam deters flying insects. It does not measure how much groundwater is available to the roots of trees near the hottest zones. The 11-per-cent-per-degree biodiversity decline is a correlation across sites, not a controlled experiment. What it tells you is what the local forest has settled into after long exposure to elevated temperature. It does not tell you exactly how a cooler forest would respond if you slowly warmed it up.

Somewhere below the forest floor, rainwater from years ago is working its way up through a fault, warming as it climbs, and delivering the last of that heat to a tributary of the Amazon at 86 degrees Celsius. The frog that falls in tomorrow will meet water that fell as rain before the researcher watching it was born.